Kawabata T, Orlandi M A, Takahashi H, Nishioka N, Okajima Y, Watanabe M, Takaishi M
Faculty of Education, Kobe University, Japan.
Health Educ Res. 1992 Sep;7(3):437-42. doi: 10.1093/her/7.3.437.
Eighty-eight second grade students of a senior high school in Saitama prefecture in Japan participated in a prospective study to predict cigarette smoking behavior 3.5 years later. Predictor variables include sex, knowledge, beliefs and attitudes toward smoking, previous smoking behavior, and smoking behavior of their families. Stepwise discriminant analyses revealed that 90% of the smokers and 65% of the non-smokers were correctly classified. In this model, previous smoking behavior proved to be the best predictor. Attitude toward adult male's smoking, sex and smoking behavior of subjects' families were also related. These four variables explained 35% of the variance in smoking behavior. As for stepwise discriminant analyses among those who had not smoked at baseline, 78% of the smokers and 76% of the non-smokers were correctly classified. Attitude toward adult male's smoking, sex, knowledge about long-term effects of cigarette smoking and smoking behavior of their families entered the model in this order. These four variables explained 37% of the variance. Implications of this study for smoking prevention programs in Japan are discussed.
日本埼玉县一所高中的88名二年级学生参与了一项前瞻性研究,以预测3.5年后的吸烟行为。预测变量包括性别、对吸烟的知识、信念和态度、以前的吸烟行为以及其家人的吸烟行为。逐步判别分析显示,90%的吸烟者和65%的非吸烟者被正确分类。在这个模型中,以前的吸烟行为被证明是最好的预测因素。对成年男性吸烟的态度、性别以及受试者家人的吸烟行为也有关系。这四个变量解释了吸烟行为中35%的方差。至于在基线时不吸烟的人群中进行的逐步判别分析,78%的吸烟者和76%的非吸烟者被正确分类。对成年男性吸烟的态度、性别、对吸烟长期影响的知识以及其家人的吸烟行为按此顺序进入模型。这四个变量解释了37%的方差。本文讨论了该研究对日本吸烟预防项目的启示。